Chapter 1 the study of collocations

3.5.2  Use of Collocations in the Essays 
 
 
The subjects' performance on the free composition task served not only 
as a measurement of the subjects' writing proficiency in English (see above), but 
 
226

---

also as a measurement of their free production of collocations.  Test papers in 
which no composition was given, either because of the particular subject's 
inadequate level of English or because of lack of time, interest etc., were not 
included in the study.  Thus, there were 275 complete test papers: 91 complete 
test papers in Group 1, 94 in Group 2, and 90 in Group 3. 
 
The essays were then analysed with regard to the collocations they 
contained.  The students' production of the 37 different collocation types as 
these are operationalised in this study (see Chapter 1) was recorded as 
frequency data.  Where the students provided a correct collocation they were 
marked as having used a token of the particular type in which the collocation 
belonged.  Misspelt collocations were recorded as evidence of collocational use. 
 Each collocation found in the essays was checked against the collocations 
included in the BBI.  If the particular collocation was included in the BBI it was 
recorded as correct evidence of use of the particular collocation type (see also 
Zhang 1993).  If it was not included in the BBI it was discarded.  There were 
13.1% of rejected collocations in all three groups.  These collocations were 
mainly Adjective Noun combinations with 'big' or 'good' as the adjective.  Such 
collocations are considered 'free combinations' by the BBI writers and they are 
not listed in the BBI (Benson et al. 1986a:xxiv). 
 
The quantity of collocations found in each essay for each of the 37 types 
was also recorded as well as the percentage of the type-token ratio.  Inter-rater 
reliability of 90% was obtained for 5% of the data.